NVIDIA GEFORCE GTX 1060 6GB
Stunning value
ALL BAD JOKES about the online merger aside, we have to admit that, as far as gaming goes, this truly is the year to be a PC gamer. Just look at what we’ve seen over the last 12 months: Skylake, Broadwell-E, Pascal, and Polaris. And with Zen round the corner, enjoying those triple-A polygons dancing across the screen as you decapitate your latest foe has never been so much fun. But the GTX 1080 and 1070 only tell part of the story. The true battle will be at the mid-range: the 1080p and 1440p cards; the grunts of both graphics companies’ armies. The flagships bring the prestige, but it’s the midrange that wins the war.
Anyway, here we have Nvidia’s latest Founder’s Edition graphics processor, in the form of the GTX 1060. Featuring 1,280 CUDA cores, 80 texture units, 6GB of GDDR5 at 8GB/s on a 192-bit bus, and nearly 4.4 billion transistors on the 16nm FinFET manufacturing process, GP106 is set to be a doozy. Like its Pascalian brothers, the GTX 1060 is designed for efficiency. And through that efficiency, Nvidia’s engineers have managed to unlock some scarily high clock frequencies on the GPU core. Nvidia will tell you that, at stock, this card’s maximum boost clock will operate at 1,708MHz, a full 530MHz higher than last generation’s GTX 960. However, that’s not entirely true. Thanks to GPU boost’s natural overclocking tendencies, at stock our card fluctuated between 1,835 and 1,860MHz during testing.
At 1080p, performance is spectacular, clocking an average frame rate of 38fps in TotalWar:Attila, 66fps in FarCryPrimal, and 60fps in TheDivision, in contrast to the GeForce GTX 980, which scored 40, 65, and 59 respectively. So, this card certainly competes with its older generation’s everyman’s flagship. It’s a similar story at 1440p, with the GTX 1060 scraping in wins or draws in all titles bar Riseof theTomb Raider, due in part to that increased VRAM capacity, and new and improved 10-bit color compression, employed to ensure the 192-bit bus isn’t a bottleneck. CLOCKING UP What’s scary, though, isn’t the stock performance, but what it can do once you delve into the overclocking settings. At the time of testing, overvolting control wasn’t fully unlocked, but we almost didn’t need it. Simply by ramping the power target up, and tweaking the GPU clock offset, we achieved an astronomical 2,151MHz on the boost clock, and an incredible 11,745 points in Fire Strike; 786 higher than at stock. With overvolting and more power-abundant aftermarket solutions, this card will be an incredibly entertaining overclocking experience, and no doubt will be pushing that 12,000-plus margin in Fire Strike well within the regions of an AMD Nano or GTX 980 OC edition. Coupled with the fact that it comes in at $300 for the Founder’s Edition, and all of a sudden Nvidia is really keeping the ball rolling by tapping into the upper echelons of the mid-range market.
Is it better than the AMD Radeon RX 480 in terms of price to performance? Almost. GP106 has proven its might, and continues to show that Pascal is not an architecture to be taken for granted. Although the Founder’s Edition still comes with a meaty price tag, and there’s no opportunity to run SLI with this card later down the line, the GTX 1060 runs cool, it runs quiet, and it’s more than capable of powering any 60fps AAA gaming experience at 1080p without a sweat.